Regardless of your political persuasion it is hard for anyone to deny the critical role that big data played in helping Barack Obama get elected for a second term. Several post-election news stories have talked about how a group of “number crunchers” in a windowless room known as “the cave” at Obama Campaign Headquarters in Chicago mined data to drive nearly every decision made throughout the two-year long campaign. From determining which media buys to make to predicting voter turnout, they measured every aspect of the campaign or as Campaign Director Jim Messina told Time “assumptions were rarely left in place without the numbers to back it up.”
Here in the battleground state of Ohio, we know firsthand about the data machine that was the Obama re-election campaign. A day after the election, a friend shared a story about a visit from an Obama campaign volunteer in the final days of the campaign. The volunteer stopped by my friend’s home to confirm whether the campaign could count on his vote. The volunteer checked his list and asked to speak to my friend’s wife and asked her whether he could count on her vote.
And then the volunteer inquired about my friend’s 22 year-old daughter, Chelsea, who still lived at home but wasn’t there at the time. Because she wasn’t home, the volunteer said someone would stop by the next day to check in with her. It turns out it was Chelsea whom the volunteer in the field was looking for. The Obama campaign data machine had determined that Chelsea, a single 22-year-old woman, was part of a key target demographic and that she was more likely to vote if she received a personal request like this from a volunteer.
So that volunteer, with Chelsea’s name on his list, was determined to have a conversation with her to make sure she went to the polls and cast her vote for Obama. The data analysts in Chicago had used determining factors such as age, sex, race, voting record to determine who were the priority targets and had deployed the volunteers in the field to find these targets and personally encourage their vote. This strategy proved highly effective when the results came in late Tuesday night. The data-driven decision making that took place in the campaign war rooms helped Obama win key battleground states and remain in the presidency.
So what can we learn from this strategy that can be applied toward using big data to drive education improvement?
- First and foremost we must organize the data. The campaign initially found that there were too many disconnected databases so they took the time and spent the necessary resources to connect the various databases to create a “megafile” from which to create comprehensive profiles of voters. This is a common problem in education – too many databases that must be connected in order to paint a complete and holistic picture of a student’s journey. In Cincinnati, the Learning Partner Dashboard was created to serve as a data aggregator and connect provider databases with the district data warehouse. Strive is working to scale this type of “megafile” database as it develops the Student Success Dashboard tool.
- Next, the data must be used to conduct small tests of change and use data for real-time continuous improvement. As an example, the campaign ran tests to determine which types of voters were persuaded by which types of appeals and made real time decisions based on what it was learning in real time. Using data in this way to drive continuous improvement is core to the collaborative action process espoused by Strive. The process requires data analysts to work together with educators and use the data to make decisions, often through small tests of change, about what individual students need to help them be successful. If a campaign can use data in this way to create a comprehensive voter profile and predict voter behavior, why shouldn’t a teacher be able to use data in the same way to create a comprehensive student profile and use that data to predict what that student will need to excel?
- Finally, ensure that data determines how resources are allocated. It’s clear that data was used throughout the campaign to target resources for the most efficient and effective uses. Examples include the strategies the campaign employed around buying advertising and where they needed to deploy field resources. In education we must use the data to determine which strategies and interventions have the greatest impact and allocate resources accordingly. This is the only way we will achieve a greater social return on investment.
The way in which the Obama campaign used big data to drive strategy is something that will be studied, published about and emulated in future campaigns. Let’s hope that the next big story is about how big data is being used to transform education. In a recently published report from Brookings, Darrell M. West ponders the learning environment of the future in which technology enables instant feedback and the teacher becomes a “data scientist.” Why must we wait for the future to achieve this? Big data is all around us. It was used to elect a President. Now, let’s put the systems and infrastructure in place so that we can use big data to ensure every student achieves, cradle to career.
Jennifer Blatz serves as Deputy Director for StriveTogether, a subsidiary of KnowledgeWorks, where she oversees StriveTogether’s strategy for accelerating the progress of cradle to career partnerships. This includes providing strategic guidance and oversight to the Cradle to Career Network, as well as StriveTogether’s impact and evaluation, knowledge management and strategic assistance functions.